What the evidence supports about practice, reflection, and the AI-augmented case method

The research community calls this the knowing-doing gap: students and professionals often understand what good judgment looks like and still fail to produce it under pressure.1Ahmadi, A., & Vogel, B. (2023).Knowing but not enacting leadership: Navigating the leadership knowing-doing gap in leveraging leadership development.Academy of Management Learning & Education, 22(3), 507–530.2Grant, R. M., & Baden-Fuller, C. (2018).How to develop strategic management competency: Reconsidering the learning goals and knowledge requirements of the core strategy course.Academy of Management Learning & Education, 17(3), 322–338. The gap is cognitive, affective, and behavioral — and what the evidence shows most clearly is that people close it through reflection and feedback on their own performance, not through additional concept-study. Students need to enact concepts in ambiguous, social, consequential situations. Concepts on their own do not get them there.

A meta-analysis of 145 studies on simulation-based learning across higher education found a large positive effect on complex-skill development, robust across medical, teacher, counseling, and management settings.3Chernikova, O., Heitzmann, N., Stadler, M., Holzberger, D., Seidel, T., & Fischer, F. (2020).Simulation-based learning in higher education: A meta-analysis.Review of Educational Research, 90(4), 499–541. A systematic review of business simulations specifically — 57 studies — similarly observes gains in knowledge, judgment, and interactive skill.4Faisal, N., Chadhar, M., Goriss-Hunter, A., & Stranieri, A. (2022).Business simulation games in higher education: A systematic review of empirical research.Human Behavior and Emerging Technologies, 2022, Article 1578791.

The experiential-learning tradition long ago argued that thinking while acting is a different cognitive activity than thinking about what one just did.5Schön, D. A. (1983).The reflective practitioner: How professionals think in action.Basic Books. Moving students from passive intake to acted experience changes what they can do, not just what they can recall.

For the past century, the case method has rested on a fixed object: a written narrative students read, prepare, and discuss. That format taught generations of professionals how to structure a problem, defend a position, and listen to a peer’s reasoning. It still does. But the format itself — a PDF, distributed in advance, debated in class — has not meaningfully evolved since well before students could ask a chatbot to summarize the protagonist’s options.

Case-based simulations are the natural next form. They keep what the case method does well — a real situation, competing stakeholders, choices with real tradeoffs — and add the layer the PDF cannot provide: a counterparty that pushes back in real time.

That live experience also produces something the static case cannot: richer raw material for the reflection that follows. In a recent study set in a senior strategy capstone, each student wrote two papers in the same course — one analyzing a traditional case, one analyzing a simulation they had completed themselves. The simulation papers scored higher on every dimension of a critical-thinking rubric, and the more deeply students engaged with the simulation, the higher their papers scored.6Samaras, S. A., Adkins, C. L., & White, C. D. (2022).Developing critical thinking skills: Simulations vs. cases.Journal of Education for Business, 97(4), 270–276. Students think more carefully about a situation they had to act inside.

The two formats are complements, not substitutes — but the static document has reached the limit of what it can do alone.

A common assumption is that the more realistic a simulation feels, the better it teaches. The evidence does not support that. What matters is whether the cues a learner needs to notice — the missed question, the conflicting incentive, the unspoken concern — are made salient enough to be noticed.

An influential management-education paper on simulation puts it this way: cognitive fidelity beats mundane realism.7Salas, E., Wildman, J. L., & Piccolo, R. F. (2009).Using simulation-based training to enhance management education.Academy of Management Learning & Education, 8(4), 559–573. A simulation works when it reproduces the cues, tradeoffs, and interpersonal pressures students will actually face — the surface details matter much less. A second principle from the same line of work: simulations collapse time. Students can run multiple decision episodes and receive rapid feedback in a context where real-world feedback is typically delayed, noisy, and confounded by everything else happening around the decision.

A simulation is an experience. Experience alone is not enough. A meta-analysis of debriefs across military, medical, aviation, and organizational settings found that structured, facilitated debriefs reliably improve subsequent performance — and that unstructured “what did we learn?” conversations don’t.8Tannenbaum, S. I., & Cerasoli, C. P. (2013).Do team and individual debriefs enhance performance? A meta-analysis.Human Factors, 55(1), 231–245. A three-arm experiment in management education sharpens the same finding: only the condition that paired action with structured dialogue and reflection produced gains in observed skill — lecture plus dialogue and reflection without an experiential layer didn’t differ meaningfully from lecture alone.9Cajiao, J., & Burke, M. J. (2016).How instructional methods influence skill development in management education.Academy of Management Learning & Education, 15(3), 508–524.

The feedback that powers a debrief rests on three questions: where am I going, how am I doing, where to next?10Hattie, J., & Timperley, H. (2007).The power of feedback.Review of Educational Research, 77(1), 81–112. Without a structured debrief that reconstructs what happened and connects choices to concepts, students often leave an exercise believing they handled it well, while the transcript tells a different story. We treat the debrief as the design center, not a finishing flourish.

One of the most counter-intuitive findings in the management-simulation literature is that gains in students’ critical-thinking scores do not consistently predict their performance in the simulations themselves.11Lovelace, K. J., Eggers, F., & Dyck, L. R. (2016).I do and I understand: Assessing the utility of web-based management simulations to develop critical thinking skills.Academy of Management Learning & Education, 15(1), 100–121. Any serious assessment of learning has to capture the reasoning behind a student’s choices, not just the outcome they produced.

The arrival of generative AI made the same point more urgent. A systematic review of academic-integrity research in AI-infused environments documents a field-wide pivot toward process-based, oral, multi-stage, and simulation-style tasks as the remaining defensible options when the line between original work and AI-assisted work has blurred.12Ncube, P. D. N., Dzvapatsva, G. P., Matobobo, C., & Ranga, M. M. (2025).Redefining student assessment in AI-infused learning environments: A systematic review of challenges and strategies for academic integrity.AI and Ethics, 6, Article 68. Static written submissions are now weak evidence of authorship; process evidence is what holds up.

This is where the Crucible Cases platform earns its place. Reflection prompts inside our simulations capture a student’s reasoning before they see the consequences of a decision. Instructor dashboards track decisions, missteps, and recoveries in real time. The transcript of how a student moved through the case — moments of confusion, course-corrections, revised judgment — is often more diagnostic than the final outcome.